I had hoped to get this article in (and out) earlier, but somehow I must have ended up on the Easter bunny’s naughty list and couldn’t find my presents in time. Nonetheless, Grand Prix São Paulo provided a lot of additional data, and I got almost all of it. Thanks to our man on the ground, Frank Karsten, I was able to figure out which decks 1,374 of São Paulo’s 1,397 competitors chose to play.

1,397 players made São Paulo the third largest Modern tournament of the year so far, and it concluded less than 12 days before the start of the most highly anticipated, high-profile, and high-stakes Modern tournament of the semester. This weekend, you can watch Magic’s biggest stars in action, live from Mythic Championship II in London. Tune in to watch GOATs battle with Gods and see the cream of the Ahn-Crop rise to the top. Meanwhile, read on to find out what happened. As the TV voice-over would put it: “Previously, in Modern…”

São Paulo Metagame Breakdown

A breakdown of the most played Modern decks at GP Sao Paulo.

Next in line were: 18 Vizier devotees, 17 Jund jugglers, 17 red-based Prison guards, 16 Eldrazi and Taxes collectors, 16 Jeskai controllers, 14 Hollow Ones, 14 Storm watchers, 13 Shadow Zookeepers, and 12 Abzan Midrange runners. Each one of these archetypes accounted for 1.29% to 0.86% of the field.

Meanwhile, a solid 14.24% share of the player base went with some more fringe strategy, of which there were about 60 different types: from 8-Rack to 8-Whack, from Elves to Faeries to Goblins and even Werewolves, from Bridgevine to Living End to Electro-End to something best described as Hollow Electro-Endvine.

So some aspects stayed very much the same as always, and that includes the almost obscene variety of decks in use in Modern. Also unchanged remained the personnel occupying the pole positions. São Paulo’s three most popular decks had already been the three most popular decks at three of the previous four Modern GPs. Izzet Phoenix once again perched atop the highest spire too.

But Izzet Phoenix failed to cross the 10% threshold, both in São Paulo as well as two weeks prior at Grand Prix Calgary. This may indicate the beginning of a trend, but it might as well be meaningless fluctuation. For some definite news, we have to look further down.

One story concerns Humans, which has been on the rise lately. Humans hovered below 4% at GP Portland in December until GP Tampa Bay in mid-March. Then, in Bilbao and Calgary, its metagame share grew past 4% and now the deck claimed 5.6%.

Another story has multiple angles and all of them involve, to some degree, Kaya, Orzhov Usurper. The card has been popping up in more and more Lantern decks, and it also appeared in a number of São Paulo’s Rock-style midrange builds. Although straight black-green remained the most popular variant and Jund stayed ahead too, Abzan gained a lot of followers. More importantly, Esper Control usurped the position formerly held by Jeskai.

Day 1 of Grand Prix São Paulo featured 21 Esper and 16 Jeskai Control players. Three Esper players advanced to the second day compared to only one Jeskai holdout. None of these numbers make for an impressive conversion rate, and there’s no breakout performance to speak of either. Though, one Esper aficionado went 11-4:

Esper Control

Bruno Silva Rocha

3 Field of Ruin
4 Polluted Delta
2 Hallowed Fountain
2 Glacial Fortress
1 Creeping Tar Pit
4 Flooded Strand
1 Godless Shrine
2 Watery Grave
1 Celestial Colonnade
3 Island
1 Swamp
1 Plains
4 Snapcaster Mage
1 Kaya, Orzhov Usurper
2 Jace, the Mind Sculptor
3 Teferi, Hero of Dominaria
1 Settle the Wreckage
1 Supreme Verdict
4 Path to Exile
1 Cast Down
3 Fatal Push
3 Nihil Spellbomb
3 Esper Charm
3 Cryptic Command
2 Logic Knot
2 Negate
2 Opt

Sideboard
2 Kalitas, Traitor of Ghet
2 Collective Brutality
2 Thoughtseize
1 Dispel
2 Unmoored Ego
1 Detention Sphere
1 Engineered Explosives
2 Celestial Purge
2 Surgical Extraction

Recently, I did a little comparative analysis of White-Blue Control decks and learned that:

Esper addresses all of these points and more. Nihil Spellbomb replaces itself and doesn’t conflict with the use of Snapcaster Mage or Logic Knot. Fatal Push looks superior to the Oust and Condemn that fill the additional spot removal slots in the more successful white-blue decks. As far as one-shot card draw goes, Esper Charm likely is an upgrade over Hieroglyphic Illumination, especially when it comes to inducing flashbacks. And virtually all Esper players agreed on Kaya. Other people made it to Day 2 with two copies in their main deck.

São Paulo Win Rates

The following includes everything—Day 1 and Day 2, as well as mirror matches. This is how the major decks performed against the full field:

Overall performance of major decks at GP Sao Paulo.

We should note that, compared to previous results, Affinity’s performance constitutes a big outlier. Hardened Scales, on the other hand, has had similar results before. All in all, it was a good weekend for the robots.

The sample for Ad Nauseam is too small to pass a simple test of statistical significance, and the same is true for all the records formed by both a double-digit number of wins and a double-digit number of losses. We might as well treat their win rates as 50%.

Nevertheless it is curious that Whir did worse even than the notoriously bad Burn. This matches a wealth of anecdotes, all of which suggest that the heyday of Whir may be over. Most notably, Louis Deltour, who reached the finals of Grand Prix Bilbao with the deck himself, said that people had started to come prepared for it. If such preparation includes enough Shatterstorm, Shattering Spree, and similar, there isn’t much Whir can do about that.

Humans spiked on Day 2 with a spectacular win rate of 68.1%. But that’s based on a scant 91 matches. Across the full two days, Humans completed 567 matches and won a much more reasonable share of 53.3% of them.

Izzet Phoenix took up 9.1% of the field on Day 1 and 12.8% on Day 2, so it’s the prime candidate to investigate the influence of mirror matches on overall win rates. There are arguments for and against their inclusion. Either way, the effect is minimal. Even the mighty Phoenix only fought 38 such matches. If we disregarded them in the calculation, the deck’s win rate would soar… all the way from 51.7% up to 51.9%.

São Paulo’s Significant Matchup Results

The chance to win a match in a matchup that’s fifty-fifty is 50%. The chance to go 2-0 in such a matchup is 25%. The chance to go 3-0 is 12.5% and the chance to go 4-0 is 6.25%. None of those are commonly considered to bear statistical significance. It is customary to deem a result significant if the chance for something this lopsided or more lopsided to occur in an even matchup comes in below 5%. A record of 5-0 qualifies, as does a 7-1, a 9-2, and so on.

I scoured São Paulo’s results for everything that matches the criteria and found the following:

As you can see, very few of the mainstream decks beat each other at a convincing rate. They are the mainstream for a reason, after all. Mostly, the more extreme matchups featured one of the more extreme strategies. Possibly the biggest exception proved Izzet Phoenix’s 66.7% win rate versus Hardened Scales. It is notable too because it closely resembles how the matchup played out at Grand Prix Bilbao. Back then, Izzet Phoenix had beat Hardened Scales 70.5% of the time.

Further corroborations: Humans had won 78.6% of matches against Burn in Bilbao, and the deck won 70.3% of matches against Burn in São Paulo. Burn had also lost horribly to Eldrazi and Taxes then and loses badly to the kindred Eldrazi now.

I’m way past my deadline, but there is additional data available. Once again, you can find all of it, all 4,908 completed matches, on this handy Google sheet. Enjoy—and enjoy the big show this weekend.